gcc/

2008-08-05  H.J. Lu  <hongjiu.lu@intel.com>

	PR target/36992
	* config/i386/emmintrin.h (_mm_move_epi64): Use
	__builtin_ia32_movq128.

	* config/i386/i386.c (ix86_builtins): Add IX86_BUILTIN_MOVQ128.
	(bdesc_args): Add IX86_BUILTIN_MOVQ128.

	* config/i386/sse.md (sse2_movq128): New.
	(*sse2_movq128): Likewise.

	* doc/extend.texi: Document __builtin_ia32_movq128.

gcc/testsuite/

2008-08-04  H.J. Lu  <hongjiu.lu@intel.com>

	PR target/36992
	* gcc.target/i386/pr36992-1.c: Scan movq.
	* gcc.target/i386/pr36992-2.c: Likewise.

	* gcc.target/i386/pr36992-3.c: New.


git-svn-id: https://gcc.gnu.org/svn/gcc/branches/ix86/avx@138734 138bc75d-0d04-0410-961f-82ee72b054a4

9 files changed, 64 insertions, 3 deletions

diff --git a/gcc/ChangeLog.avx b/gcc/ChangeLog.avx
index 81a1c66c856..14e4d98550e 100644
--- a/gcc/ChangeLog.avx
+++ b/gcc/ChangeLog.avx
@@ -1,3 +1,17 @@
+2008-08-05  H.J. Lu  <hongjiu.lu@intel.com>
+
+	PR target/36992
+	* config/i386/emmintrin.h (_mm_move_epi64): Use
+	__builtin_ia32_movq128.
+
+	* config/i386/i386.c (ix86_builtins): Add IX86_BUILTIN_MOVQ128.
+	(bdesc_args): Add IX86_BUILTIN_MOVQ128.
+
+	* config/i386/sse.md (sse2_movq128): New.
+	(*sse2_movq128): Likewise.
+
+	* doc/extend.texi: Document __builtin_ia32_movq128.
+
 2008-08-04  H.J. Lu  <hongjiu.lu@intel.com>
 
 	* config/i386/sse.md (*avx_shufps_1): Removed.
diff --git a/gcc/config/i386/emmintrin.h b/gcc/config/i386/emmintrin.h
index 933dcd61e63..c6590dce4d4 100644
--- a/gcc/config/i386/emmintrin.h
+++ b/gcc/config/i386/emmintrin.h
@@ -726,7 +726,7 @@ _mm_movpi64_epi64 (__m64 __A)
 extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
 _mm_move_epi64 (__m128i __A)
 {
-  return _mm_set_epi64 ((__m64)0LL, _mm_movepi64_pi64 (__A));
+  return (__m128i)__builtin_ia32_movq128 ((__v2di) __A);
 }
 
 /* Create a vector of zeros.  */
diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index d00065ca63d..3cfe3170b1b 100644
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -19116,6 +19116,8 @@ enum ix86_builtins
   IX86_BUILTIN_MOVNTPD,
   IX86_BUILTIN_MOVNTDQ,
 
+  IX86_BUILTIN_MOVQ128,
+
   /* SSE2 MMX */
   IX86_BUILTIN_MASKMOVDQU,
   IX86_BUILTIN_MOVMSKPD,
@@ -20467,6 +20469,8 @@ static const struct builtin_description bdesc_args[] =
   { OPTION_MASK_ISA_SSE2, CODE_FOR_abstf2, 0, IX86_BUILTIN_FABSQ, UNKNOWN, (int) FLOAT128_FTYPE_FLOAT128 },
   { OPTION_MASK_ISA_SSE2, CODE_FOR_copysigntf3, 0, IX86_BUILTIN_COPYSIGNQ, UNKNOWN, (int) FLOAT128_FTYPE_FLOAT128_FLOAT128 },
 
+  { OPTION_MASK_ISA_SSE, CODE_FOR_sse2_movq128, "__builtin_ia32_movq128", IX86_BUILTIN_MOVQ128, UNKNOWN, (int) V2DI_FTYPE_V2DI },
+
   /* SSE2 MMX */
   { OPTION_MASK_ISA_SSE2, CODE_FOR_mmx_addv1di3, "__builtin_ia32_paddq", IX86_BUILTIN_PADDQ, UNKNOWN, (int) V1DI_FTYPE_V1DI_V1DI },
   { OPTION_MASK_ISA_SSE2, CODE_FOR_mmx_subv1di3, "__builtin_ia32_psubq", IX86_BUILTIN_PSUBQ, UNKNOWN, (int) V1DI_FTYPE_V1DI_V1DI },
diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index 9ea8d8f5daa..9f9f5e889ec 100644
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -339,6 +339,28 @@
    (set_attr "prefix" "vex")
    (set_attr "mode" "<MODE>")])
 
+(define_expand "sse2_movq128"
+  [(set (match_operand:V2DI 0 "register_operand" "=x")
+	(vec_concat:V2DI
+	  (vec_select:DI
+	    (match_operand:V2DI 1 "nonimmediate_operand" "x")
+	    (parallel [(const_int 0)]))
+	  (match_dup 2)))]
+  "TARGET_SSE2"
+  "operands[2] = CONST0_RTX (DImode);")
+
+(define_insn "*sse2_movq128"
+  [(set (match_operand:V2DI 0 "register_operand" "=x")
+	(vec_concat:V2DI
+	  (vec_select:DI
+	    (match_operand:V2DI 1 "nonimmediate_operand" "x")
+	    (parallel [(const_int 0)]))
+	  (match_operand:DI 2 "const0_operand" "")))]
+  "TARGET_SSE2"
+  "movq\t{%1, %0|%0, %1}"
+  [(set_attr "type" "ssemov")
+   (set_attr "mode" "TI")])
+
 (define_insn "<sse>_movup<ssemodesuffixf2c>"
   [(set (match_operand:SSEMODEF2P 0 "nonimmediate_operand" "=x,m")
 	(unspec:SSEMODEF2P
diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi
index 9557ba3bad9..66fb8401c78 100644
--- a/gcc/doc/extend.texi
+++ b/gcc/doc/extend.texi
@@ -8071,6 +8071,7 @@ v2di __builtin_ia32_psrlqi128 (v2di, int)
 v8hi __builtin_ia32_psrawi128 (v8hi, int)
 v4si __builtin_ia32_psradi128 (v4si, int)
 v4si __builtin_ia32_pmaddwd128 (v8hi, v8hi)
+v2di __builtin_ia32_movq128 (v2di)
 @end smallexample
 
 The following built-in functions are available when @option{-msse3} is used.
diff --git a/gcc/testsuite/ChangeLog.avx b/gcc/testsuite/ChangeLog.avx
index 6481e2be699..5d718b8f3c7 100644
--- a/gcc/testsuite/ChangeLog.avx
+++ b/gcc/testsuite/ChangeLog.avx
@@ -1,3 +1,11 @@
+2008-08-04  H.J. Lu  <hongjiu.lu@intel.com>
+
+	PR target/36992
+	* gcc.target/i386/pr36992-1.c: Scan movq.
+	* gcc.target/i386/pr36992-2.c: Likewise.
+
+	* gcc.target/i386/pr36992-3.c: New.
+
 2008-07-12  H.J. Lu  <hongjiu.lu@intel.com>
 
 	* gcc.dg/compat/vector-1b_main.c: New.
diff --git a/gcc/testsuite/gcc.target/i386/pr36992-1.c b/gcc/testsuite/gcc.target/i386/pr36992-1.c
index aad6f7cd14d..7cd24cccf3e 100644
--- a/gcc/testsuite/gcc.target/i386/pr36992-1.c
+++ b/gcc/testsuite/gcc.target/i386/pr36992-1.c
@@ -9,4 +9,4 @@ test (__m128i b)
   return _mm_move_epi64 (b);
 }
 
-/* { dg-final { scan-assembler-times "mov\[qd\]\[ \\t\]+.*%xmm" 1 } } */
+/* { dg-final { scan-assembler-times "movq\[ \\t\]+.*%xmm" 1 } } */
diff --git a/gcc/testsuite/gcc.target/i386/pr36992-2.c b/gcc/testsuite/gcc.target/i386/pr36992-2.c
index eb9c3a28fee..8c9089460f0 100644
--- a/gcc/testsuite/gcc.target/i386/pr36992-2.c
+++ b/gcc/testsuite/gcc.target/i386/pr36992-2.c
@@ -9,4 +9,4 @@ test (__m128i b)
   return _mm_move_epi64 (b);
 }
 
-/* { dg-final { scan-assembler-not "%mm" } } */
+/* { dg-final { scan-assembler-times "movq\[ \\t\]+.*%xmm" 1 } } */
diff --git a/gcc/testsuite/gcc.target/i386/pr36992-3.c b/gcc/testsuite/gcc.target/i386/pr36992-3.c
new file mode 100644
index 00000000000..17696a5b276
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr36992-3.c
@@ -0,0 +1,12 @@
+/* { dg-do compile }
+/* { dg-options "-O2 -msse4" } */
+
+#include <emmintrin.h>
+
+__m128i
+test (__m128i b)
+{
+  return _mm_move_epi64 (b);
+}
+
+/* { dg-final { scan-assembler-times "movq\[ \\t\]+.*%xmm" 1 } } */